Development Status¶
Alpha. ProbOS implements one governed Cognitive Mesh with an experimental federation transport. The getting-started path is Beta: an automated test runs the Quickstart from probos setup on.
Snapshot (October 2026)¶
Measured on main on 2026-10-07. The agent figures come from booting the shipped config/system.yaml with the built-in mock LLM client; a smaller configuration, such as the one probos setup writes, boots fewer agents.
| Metric | Value |
|---|---|
Python tests (pytest --collect-only) |
39,707 |
| Vitest test files | 357 (HXI) + 10 (desktop host) |
Python modules in src/probos |
993 (~344,000 lines) |
| Agents at boot | 81, in 51 pools |
| Sovereign crew agents | 15, in 6 departments |
| Pool groups | 9 (Bridge, Core Systems, Engineering, Medical, Operations, Science, Security, Self-Modification, Utility) |
| Shell slash commands | 61 |
| API router modules | 67 |
Sections in config/system.yaml |
181 |
| Architecture decisions | AD numbers past AD-1300 |
| Latest tag | v0.4.0-phase29c (March 2026); no release since — PyPI packaging is tracked in #1053 |
Eras¶
| Era | Highlights |
|---|---|
| I — Genesis (Phases 1–9) | Substrate, mesh, consensus, cognitive core, shell, federation transport |
| II — Emergence (Phases 10–21) | Self-modification, skills, LLM tiers, dreaming, persistent knowledge, Shapley trust |
| III — Product (Phases 22–29) | HXI, Discord adapter, bundled agents, medical and science teams |
| IV — Evolution (Phase 30+) | Standing orders, Ward Room, sovereign DID identity, ontology, Cognitive JIT, self-regulation, emergence metrics |
| V — Unification | Knowledge graph, brain-enhancement work, commercial-overlay seam |
| VI — Commercial Bridge (May 2026) | Extension-point registry, NATS cognitive-chain pipeline, contagion firewall, voice substrate |
Since June 2026 the work has been organized as programs rather than eras:
| Program | What landed |
|---|---|
| Agentic execution and tools | A multi-turn agentic loop with tool search and delegation; MCP servers with per-agent and per-tool authorization; governed Python execution in per-agent workspaces; office-document agents; a browser tool and browser workstation |
| Crew execution and approvals | Durable crew sessions with recovery and verified finalization; work items agents can discover and claim; an approvals center; delegated approvals with chain-of-command routing and decision pre-clearance |
| Composable cognition | The cognitive spine and organ contracts, with attention and dreaming as the first organs (design) |
| Onboarding | probos setup provider wizard, probos doctor, and a tested Quickstart |
| Identity and federation | Ed25519 key binding for the ship's did:probos identity, signed federation envelopes, and authenticated peer admission proven between two separately started nodes (AD-1196–1198) |
| Platform maturity | The AD-1270 program: exercised capabilities, robust seams, modular owners, fast gates and truthful docs |
Current Priorities¶
From the Roadmap (2026-09-07):
- Finish the dependable mesh — AD-1270 / #1324.
- Measure benefit as paths become ready — the Sigma ablation (#1064) and Ship Trials (#1123).
- Bring human contribution and independent adoption forward — typed human claims (#1136) and a PyPI release (#1053).
- Authenticate before wider federation — fleet policy (#479) follows the two-node authentication now in place.
- Keep extensions tied to outcomes — elastic-team trials (#1337) and governed self-maintenance (#1352).
Nooplex Readiness¶
From the Nooplex Readiness Map, which is the authority for these claims:
| Tier | Name | Status | Permitted claim |
|---|---|---|---|
| A | Dependable Cognitive Mesh | planned |
"ProbOS is an alpha implementation of a governed Cognitive Mesh." |
| B | Secure Multi-Mesh | planned |
"ProbOS has an experimental federation transport." |
| C | Nooplex Core Fabric | not-started |
"ProbOS provides local foundations for a future Nooplex Core." |
| D | Emergence Research | research |
"ProbOS is an experimental platform for testing the Nooplex hypothesis." |
Live GitHub issues and their evidence receipts are the completion authority; PROGRESS.md and DECISIONS.md hold the narrative and decision logs.